In order to perform tasks of electronic devices well or better, the electronic devices need to be connected with each other with cable or communication line to form a network, so as to set up communication relationship, and to exchange information to collaborate their work. Functionally, an electronic device in the network can be categorized either as a termination device or a forwarding device. The termination device is used for sending and receiving information. Further, a termination device is called a source device when it plays role of sending information, or the termination device is called a destination device when it is used for receiving information. A forwarding device is an intermediate device used for forwarding information and it forwards information in the process of delivering information from source device to destination device. For instance, IP router in IP network and ATM (Asynchronous Transfer Mode) switch in ATM network are both examples of forwarding devices.
The above means that an electronic device may play three roles in the process of data transmission: the source device, the forwarding device and the destination device. In fact, the same electronic device can play different roles in different circumstances. For example, sometimes a forwarding device just receives data and doesn't forward the data out. It works as destination device at this moment. To make the description easy, it limits an electronic device to play a single role below. That means one electronic device must be only one of the roles of the source device, the forwarding device, and the destination device. In fact, the limitation does not affect generality of the method discussed. In order to communicate with each other after the electronic devices form a network, it is necessary to develop a coding method which is used to define an identifier for each electronic device, or they can't communicate. That identifier is called address of the electronic device.
IP address in IP network and VPI/VCI address in ATM network are addresses of electronic device commonly used. In the IP network, each electronic device is assigned an explicit and constant address. That is to say, coding method of absolute address is used and the address is called absolute address. The characteristic of the coding method of absolute address is that the destination device is one-variable function of absolute address. No matter when and where, once the absolute address is given, the destination device is determined. The relationship is shown as follows:Dd=f(Aa)Where,                Aa refers to the absolute address;        Dd refers to the destination device located by Aa; and        f refers to an abstract function, denoting the relationship between Dd and Aa.        
How about the performance of the coding method for network address can be appraised by the performance of the network established on this kind of network address, such as security, complexity, scalability, and so on. Problems of the insecurity, complexity, and un-scalability in IP network are mainly due to the coding method for network address in IP network, which is illustrated as follows.                Firstly, IP network is insecure. Three main reasons can be accounted for that.        (1) The destination address in a data packet must be visible directly to forwarding device (i.e. IP router). It can't be encrypted, or routers will not identify the destination address and can't forward data.        (2) Since length of IP address is fixed and limited, all possible IP addresses can be enumerated. That means all IP addresses can be searched to obtain, which may cause insecurity.        (3) Since a fixed address is assigned for an electronic device, once the address is acquired, anyone can access the electronic device from anywhere.        Secondly, the architecture of IP network is complex. An IP address is just the identification of an electronic device, and it doesn't contain routing information directly which is stored in a forwarding device for IP network. As the forwarding device is used by lots of users simultaneously, it is very complicated for the forwarding device to obtain and maintain dynamic routing information. Furthermore, it consumes a huge quantity of computing resources to access enormous routing information in a router, which makes the forwarding device in IP network need large storage space and high computing capability. Hence, IP router and the whole network as a result are high cost, and there may exist high speed bottleneck in IP network, restricting the application of the network.        Thirdly, the IP network architecture has poor scalability and adaptability. Reasons are as follows:        (1) The duplication of IP addresses is strictly not allowed. That means all addresses of electronic devices in the network can't be the same, so that it has to solve the duplication of names in the whole network range. A distribution system of network address is set up to solve the problem. That is to say, we can not expand the network arbitrarily and have to raise the procedure to apply an address before the expansion of the network, which is not convenient.        (2) The range of IP addresses is fixed. On the one hand, it has problem of address starvation for a large network. On the other hand, the long IP address is wasteful and brings an overhead relative big for small network.        